A smooth penalty-based sample average approximation method for stochastic complementarity problems

Sample average approximation method is one of the effective methods in the stochastic optimization. A smooth penalty-based sample average approximation method for stochastic nonlinear complementarity problems is presented in this paper. Based on a smooth penalty function, a reformulation is proposed for the equivalent problem of EV formulation for stochastic complementary problems and it is proven that its solutions are existent under some mild assumptions. An implementable sample average approximation method for the reformulation is further established and its convergence is analyzed. The numerical results for some test examples are reported at last to show efficiency of the proposed method.